[Bf-committers] Seamless texture mapping

Vincent Nivoliers vincent.nivoliers at loria.fr
Tue Jun 1 13:00:52 CEST 2010

   I surely understand that you have lots of other things to do. We have 
indeed some code to share, and that was the objective of this proposal. 
We have here a partial reproduction of mixed integer quadrangulations, 
with some modifications (mostly we don't have the final tricks of the 
paper, and we changed the expression of the minimized energy for the 
parameterization). We also have Direction Field Design that we use to 
generate the direction field for MIQ instead of theirs (which presumably 
doesn't handle high frequencies in the geometry of the mesh very well, 
but that's a supposition). The first author of our paper also 
implemented Quadcover. All this was part of the first "grid preserving 
parameterizations" part I was talking about in the paper. All this code 
was supposed to be present in the clean graphite version I was talking 
about. However, this is graphite code, based on its halfedge mesh data 
structure and attribute manager, I don't have a clue about how difficult 
it would be to transfer it to blender.
   Thanks for your concern,
> Hi Vincent,
> This looks interesting, though I personally don't have time to help
> out. Unfortunately even if we get features handed on a golden plate,
> integration still takes time, but let's see if there are other
> developers who would like to pick this up.
> With 2.5 it is also possible to distribute this as an add-on,
> depending on how closely it needs to integrate. We also currently have
> a summer of code project to implement mixed integer quadrangulation,
> perhaps some code sharing would be possible if it is integrated
> natively.
> Brecht.
> On Mon, May 31, 2010 at 1:17 PM, Vincent Nivoliers
> <vincent.nivoliers at loria.fr>  wrote:
>> Hi all,
>>    I'm contacting you to know whether you would be interested about a
>> paper I just got accepted. You can find the paper and video at
>> http://alice.loria.fr/~nivoliev/Seamless
>>    Currently the code is part of a graphite plugin, you might know the
>> software since you already worked with Bruno Lévy for LSCM, who is also
>> among the authors of this paper. It is therefore not in a very proper
>> state to be shared. Before cleaning up the code, we would like to know
>> whether you might be interested in these results. We tried to think the
>> paper as a useful contribution, however the effective users might have a
>> different opinion.
>>    At first we could try to build a clean graphite plugin for you to
>> test the thing if you are interested. Tell me from the paper if you
>> would like this. If you then feel it as worth the proper integration, we
>> would be glad to help you, by trying to deliver a more standalone
>> version of the code.
>>    Basically, there are distinct parts in the project. The first one is
>> not discussed in the paper and corresponds to grid preserving
>> parameterizations. You might have read about it, it is mostly used for
>> quad remeshing. The second part builds a texture atlas from such a
>> parameterization, and the necessary stuff for seamless mapping. The last
>> one is a texture fixing program correcting an input texture using the
>> stuff from the previous step to map it seamlessly on the atlas built
>> also in the previous step. The "integration" you see in the video is
>> just a two small scripts for importing and exporting meshes from blender
>> to graphite, so no real integration in fact.
>>    I hope you'll find it interesting, and thanks for the wonderful
>> software you develop. Most of the video was done using it (3D for the
>> animations, and VSE for the final assembly).
>> Regards,
>> --
>> Vincent Nivoliers
>> http://alice.loria.fr/~nivoliev
>> _______________________________________________
>> Bf-committers mailing list
>> Bf-committers at blender.org
>> http://lists.blender.org/mailman/listinfo/bf-committers
> _______________________________________________
> Bf-committers mailing list
> Bf-committers at blender.org
> http://lists.blender.org/mailman/listinfo/bf-committers

More information about the Bf-committers mailing list